Automatic Path Planning for Robotic Grinding and Polishing Tasks based on Point Cloud Slicing

The remanufacturing of workpieces suffers from a high manual effort, which emphasises the need for automation of such processes by robots. However, this is associated with some challenges due to the high variance in terms of individual characteristics of possible defects. Shape and surface properties deviating from the ideal virtual 3D model require approaches that allow path planning to be carried out directly on the component to be machined. This paper proposes a novel automatic path planning algorithm for robotic grinding and polishing tasks based on point clouds obtained directly by scanning the workpiece. The adaptive determination of the position and orientation of intersection planes as well as a slicing direction enables obtaining cross-sectional contours by slicing the point cloud. The algorithm generates a continuous, multidirectional path that has a meander shape precisely adapted to the workpiece surface. A qualitative analysis emphasises the robustness of the algorithm with respect to the distribution of the point cloud data and limited amount of noise.